Defining the Built Environment in the Context of Physical Activity:

Activity: Neighbourhood Opportunity Structures

Of the 46 studies, 27 (58.7%) used methodologies to gain insight into neighbourhood opportunities for physical activity. Within these studies, there were three main ways of measuring and assessing the built environment for physical activity: buffers;

administrative units; and straight-line distance.

Of the studies using these neighbourhood proxies, the majority of studies (19) used buffer-based measures (70.4%). Even within these buffer-based measures, there is no consensus on which buffer size best captures a child’s neighbourhood environment. The majority of buffer-based studies (73.7%) used either a single 800 metre (0.5 mile) or 1600 metre (1 mile) buffer around the home. Multiple buffers were used in only 3 studies. The smallest buffer size used across all buffer-based studies was 200 metres and was in a study using multiple buffers (Van Loon, Frank, Nettlefold, & Naylor, 2014). The largest buffer size used across all buffer-based studies was 2000 metres (Crawford et al., 2010; Prins et al., 2011). 8 studies (29.6%) used administrative units (i.e., division of a region) as a measure of the environment. Similar to the studies using buffer-based measures, there is a great deal of heterogeneity regarding the type of administrative unit used.

In studies using neighbourhood proxies, the outcome measure was constrained by the use of a neighbourhood proxy. In these studies, the outcome was an average daily or weekly

(a) minutes of physical activity, (b) counts per minute or epoch, or (c) steps. Because the actual locations of physical activity are unknown, these studies have to assume that all physical activity occurred within their neighbourhood proxy and use the average estimates of physical activity.

Of the 27 studies using neighbourhood proxies, only 10 used objective measures of physical activity and objective measures of the built environment exclusively. The remaining 17 studies used a combination of objective measures and subjective measures. Four studies used objective and subjective measures of physical activity alongside objective measures of the environment. Ten studies used objective measures of physical activity alongside objective and subjective measures of the environment. Three studies used objective and subjective measures of physical activity alongside objective and subjective measures of the environment.

2.3.4

Defining the Built Environment in the Context of Physical

Activity: Environmental Exposure

Of the 46 studies, 19 (41.3%) used methods to gain insight into the spaces that children used for physical activity. In other words, these studies assessed exposure to physical activity environments. Within these studies, there was one primary methodology used: combining GPS tracking with accelerometer data and integrating the data within a GIS. Despite using one main methodology across all studies, there were still methodological differences across the studies when characterizing the environment. In some studies, the GPS tracking was done alongside buffers, administrative units, and straight line

distances.

Of the studies using accelerometer-GPS data, 6 (31.6%) used additional buffers. Two of these 6 studies buffered every accelerometer-GPS point while 4 of these 6 studies used GPS-accelerometer data alongside neighbourhood proxies to characterize the

neighbourhood environment. In addition, 1 study used administrative units alongside the accelerometer-GPS tracking as a proxy for the child’s neighbourhood, and 1 study used straight-line distance to the nearest park boundary from the participants’ home address.

The remaining 11 studies (57.9%) used GPS tracking as the only way of measuring environmental exposure with simultaneous accelerometry.

The physical activity outcome measures in studies using simultaneous GPS tracking and accelerometry were diverse. Studies used a variety of outcomes ranging from bouts (the percentage of bouts, the number of bouts), METs (MET weighted MVPA, MET for each GPS point), activity counts (total number, counts per minute, mean, or the percentage of counts), the average daily/weekly number of minutes, counts, or steps, the time spent at different locations (the number of minutes, the proportion of time spent), and the probability of MVPA at each epoch.

Of the 19 studies using accelerometer-GPS data, 13 used objective measures of both physical activity and the environment exclusively. The remaining papers used subjective measures alongside objective measures. Four papers used objective and subjective measures of physical activity alongside objective measures of the environment. One study used objective measures of physical activity alongside objective and subjective measures of the environment. One study used objective and subjective measures of physical activity alongside objective and subjective measures of the environment. Overall, only 2 studies used subjective measures of the environment in addition to the GPS tracking (Table 2.3).

Table 2.3 Objectivity characteristics for each study

Measurement Characteristics of the Papers Number of Papers

Studies Using Neighbourhood Proxies 27

Objective PA; Objective Environment 10

Objective and Subjective PA; Objective Environment 4 Objective PA; Objective and Subjective Environment 10 Objective and Subjective PA; Objective and Subjective Environment 3

Studies Using GPS Monitoring 19

Objective PA; Objective Environment 13

Objective and Subjective PA; Objective Environment 4 Objective PA; Objective and Subjective Environment 1 Objective and Subjective PA; Objective and Subjective Environment 1

2.3.5

Environmental Correlates of Physical Activity

Regardless of methodology used (i.e., neighbourhood proxies versus accelerometer-GPS data), there were marginally more null relationships found than significant (both positive and negative in direction) relationships (Tables 2.4 and 2.5). Several variables had inconsistent associations, particularly measures of parks and recreation facilities. Papers using neighbourhood proxies to measure the environment not only examined different environmental correlates of physical activity, but found different significant relationships compared to papers using accelerometer-GPS data.

Table 2.4 Number of papers with significant relationships for each environmental attribute in studies using neighbourhood proxies

Objectively Measured Environmental Variables

Results Count

+ 0 -

Recreation Environment

Parks (acces/density/proximity) 5 10 1

Recreation facilities (access/density/proximity) 1 7 4

Neighbourhood Design

Accessibility index 0 0 0

Commercial density 1 1 0

Cul-de-sac density 1 0 2

Employment density 1 1 0

Land Use Mix 2 2 0

Neighbourhood type 3 3 3

Population Density 1 2 0

Residential Density 1 3 1

Street connectivity 1 3 0

Urbanicity (significant difference between groups) 2 0 1

Walkability 4 3 0 Transportation Environment Pedestrian aesthetics 2 2 1 Pedestrian amenities 3 6 0 School (distance) 0 4 4 Traffic speed/volume 4 6 0 Other Beaches 0 0 0 Farmland 0 0 0 Gardens 0 0 0 Grassland 0 0 0 Greenspace/NDVI 0 1 0 Non-recreational buildings 0 1 1 Open space 1 2 1

Other built land (e.g. playground) 1 2 0

Roads/pavements 2 0 0

2.3.6

Environmental Correlates and Physical Activity: